Hi, this may help you to detect the cause
Really interesting. Problem is, I have no 64bit Linux platform for testing at hand. And it's nothing that can be deduced from program flow and logic.
My workaround does nothing useful in theory, it just stops using the this-pointer as status variable. Now if the pointer keeps ketting corrupted, at least with tray icon off nothing bad happens. But - luquino has tested it with tray icon
on and it still works. So the read access to the pointer appears to corrupt it ? Whenever a message comes in ?

Your workaround for the tray icon pointer will be inserted in SVN version? If yes, when?
It's in trunk since Thursday (8268).
Thank you so much for your help.
You're welcome! Thank you for helping to get to the root of the problem (even if I still haven't found it).